The present invention relates to a technique for processing audio signals that represent sounds such as instrumental or vocal sounds.